LA-8336-MS
Cr5.—c
C3. ClC-l 4 REPORT COLLECTION
REPRODUCTION
COPY
Criticality Calculations
and Criticality Monitoring Studies of the
Slagging Pyrolysis Incinerator Facility
LOS ALAMOS SCIENTIFIC LABORATORYPost Office Box 1663 Los Alamos. New Mexico 87545
An Affirmative Aciion;F+al Opportunity Employet
.
Edited by
Christine L. B. Siciliano
Photocomposition by
Chris West and Joni Lee Powel
This work was supported by the US Depart-
ment of Energy, Transuranic Waste Manage-
ment Programs and Idaho Natioml Engineering
12boratory.
DISCLAIMI:R
This report was prepncd as an account of wotk sponsored by an agency of the United Slates Gc.vern.mcnt. Neither the United States Government noc any agency Ihewof, nor any or thck employees,mskes any warranly, express or implied, or assumesany Icgal Ii’’bdity or rcsponsibtity for the accuf-ac’y, rnmplcteness, or usefulness of any information, sppar~tus, product, or process disclosed, or rep.resents that its usc would not infringe privately owned rights. Reference herein to any spaitic tom.merchl produci, process, or service by trade name, trademark, manufactwer, or othezwise. does notnecessarily rnnstitute or imply its endorsement, recommendation, or favoring by the United SatesGovcmment or any agency thereof. The views and opinions of authors expressed herein do not ncc-e~rily state or reflect those of the United Stales Govunment or any agency thereof.
UNITED STATES
DEPARTMENT OF ENERGY
CONTRACT W-7405 -ENG. 36
LA-8336-MS
4.
Criticality Calculations
and Criticality Monitoring Studies of the
Slagging Pyrolysis Incinerator Facility
D. A. CloseT. E. Booth
J. T. Caldwell
!+ ~ -“- ‘- ;-
b~-:... i,=-- ...-yr ,_ ~
[ ~~ ,---- ‘~..—..
.,. -..+..—.,r
.- ... ,.. .-_,. -A -.
. . . . . ~..
. . . . . . . .. . ,:r. .-. m- ,-----. .
UC-46Issued: January 1981
. .. . . . . . . . ... ’.*. ~.g.,
—. ..- . . . . . . .
CONTENTS
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...1
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . .00......o....2
I.
H.
111.
Iv.
v.
MONTE CARLO NEUTRON CALCULATIONS OF THE CRITICALITY HAZARD IN
THE SPIFACIILITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...2
A. Section Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...2
B. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...2
C. BriefCodeDescription. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...3
D. Material Compositions and Region Temperatures . . . . . . . . . . . . . . . . . . . . 4
E. Incinerator Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...6
F. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...7
G. Discussion of Incinerator Results . . . . . . . . . . . . . . . . . . . . . . . . . ...8
H. Secondary Combustion Chamber, Slag Tap, and Holding Tank . . . . . . . . . . . . .11
I. Tundish . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...11
J. Off-Gas Collection Hopper. . . . . . . . . . . . . . . . . . . . . . . . . . . . ...12
K. Conclusions From MonteCarloStudies . . . . . . . . . . . . . . . . . . . . . . . . 12
DIRECT CRITICALITY MONITORING . . . . . . . . . . . . . . . . . . . . . . . ...13
A. Section Summary . . . . . . . . . . . . . . . . .1....... . . . . . . . . ...13
B. Passive Neutron System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...13
C. Active Neutron System.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...17
DISTINGUISHING BETWEEN FISSILE AND FERTILE MATERIALS IN NUCLEAR
WASTE ATALLSTAGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...17
A. Section Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...17
B. Active Thermal Neutron Fission Techniques . . . . . . . . . . . . . . . . . . . . . 0 17
C. Passive Neutron Assay Techniques . . . . . . . . . . . . . . . . . . . . . . . . ...19
D. SPIFacility Input Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...20
ASSAY OF PRODUCT SLAGCASTINGS . . . . . . . . . . . . . . . . . . . . . . ...20
CONCLUSIONS, RECOMMENDATIONS, ANDCOSTESTIMATES . . . . . . . . . .21
A. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..21
B. Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...21
C. Cost Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...21
ACKNOWLEDGMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..s . ...22
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..22
APPENDIX: SPI FACILITY GEOMETRY FORMCNP . . . . . . . . . . . . . . . . . . . . .23
.8
.+ I
iv
CRITICALITY CALCULATIONS AND CRITICALITY
MONITORING STUDIES OF THE SLAGGING
PYROLYSIS INCINERATOR FACILITY
by
D. A. Close, T. E. Booth, and J. T. Caldwell
ABSTRACT
We determined that the criticality hazard associated with the Slagging Pyrolysis
Incinerator (SPI) Facility proposed by Maho National Engineering Laboratory would
be minimal if a three-level criticality-hazard prevention program were implemented The
fwst strategy consists of screening all incoming wastes for fissile content. An estimate of
the average fissile content of wastes to be processed by the SPI Facility is S1 g fissile per
50 kg waste. Monte Carlo calculations indicate that under normal operating conditions
this average fissile concentration poses no criticality hazard. A tissile concentration in
the SPI Facility ten times this average value is required to produce k of about 0.4. Large
lumps are not allowed to enter the system, which effectively prevents a rapidly
developing criticality hazard. Suitable instrumentation to easily and reliably accomplish
this screening/assay is currently under development at both EG&G, Inc. and Los
Alamos Scientific Laboratory.
The second prevention level is provided by introducing a small concentration of a
neutron-absorbing compound, such as BZOJ, into the input waste stream. Our Monte
Carlo calculations showed, for instance, that including only 1.0 x 10-2 w fraction of
BzOj reduces the criticality hazard to a negligible level.
Examph: For 9.3 kg 239Pu in the drying region without BZOJ, k = 0.99. For the same
93 kg of ‘3gPu in the drying region with 1.0 x 10–2 w fraction of BZ03.
added, k = 0.24.
The third prevention level is provided by direct criticality-hazard monitoring using
sensitive neutron detectors in all regions of the facility where a signitlcant hazard has
been identified-principally the drying, pyrolysis, and slag regions. These detectors can
be operated in a passive mode or can be used with a pulsed external neutron source to
obtain a direct measure of k. These neutron detectors have been used in other
applications to measure k in the range 0.05 to 0.10. They can also be used to determine
an approximate value of the total fissile mass in the SPI Facility. On the basis of thh
data, the facility could be shut down rapidly for cleanout if these measurements indicate
an unsafe condition is developing. The criticality safety provided by the product of these
three independent measures should reduce the hazard to a negligible level. We
recommend that they be implemented in the SPI Facility. A detailed fault tree statistical
analysis should be done to quantify the safety factors associated with each of the three
levels of this criticality-hazard prevention program. Such an analypis is beyond the
scope of this report.
——.———————————————
1
INTRODUCTION
The Los Alamos Scientific Laboratory (LASL) study
of the proposed Slagging Pyrolysis Incinerator (SPI)
Facility for criticality hazard and for appropriate
criticality monitoring was initiated at the request of
EG&G, Inc., Idaho Falls, Idaho. UASL agreed to
independently study the problem of criticality hazard in
the SPI Facility and to recommend, on the basis of this
study, appropriate criticality-monitoring instrumentation
and/or procedures to minimize or lessen potential
criticality hazards.
Speciiic analyses for this study included the following.
I. A Monte Carlo neutron study was performed to ( 1)
examine the effects of neutron poisoning and various
typical matrix materials on criticality, as well as their
effects on possible criticality-monitoring systems; (2)
identify areas of the SPI Facility where criticality
hazards exist; and (3) quantify the hazards as much as
possible. Specifically, we addressed the problems as-
sociated with plutonium lumps being processed through
the system and quantified the criticality hazards as-
sociated with plating out of plutonium in various regions.
H. A study of direct criticality-monitoring techniques
and equipment included both active and passive neu-
tron-based methods, with a limited number of experimen-
tal measurements on multiplying systems. Recommenda-
tions of the appropriate technique(s) with estimates of
needed development and associated costs were included.
III. The general problem of distinguishing between
fissile and fertile materials in waste both before and atler
slagging was studied.IV. Methods for assaying product slag in
91.44 -cm-(36 -in.-) diam x 30.48 -cm-( 12-in.-) thick ingots
were studied.
L MONTE CARLO NEUTRON CALCULATIONS
OF THE CRITICALITY HAZARD IN THE SPI
FACILITY
A. Section Summary
Monte Carlo calculations were performed to de-
termine critical plutonium concentrations in various
regions of the proposed SPI Facility. The incinerator was
divided into three plutonium-containing regions: the
drying, pyrolysis, and slag regions. Calculations in-
dicated that these regions are only weakly coupled
neutronicail y; that is, when each region by itself is close
to critical, the whole incinerator remains close to critical.
The critical plutonium weight fractions for the drying,
pyrolysis, and slag forming regions are 3.0 x 10-3, 2.6 x
10-3, and 2.0 x 10-2, respectively. These weight fractions
translate into 239Pu critical masses of 9.3, 28.7, and 271
kg, respectively. In other words, the minimum critical
mass of 239Pu for the SPI Facility as a whole is 9.3 kg if
the entire 9.3 kg is placed within the drying region.
Under some operating conditions, fissile material
build-up within the SPI Facility is possible. In this
situation, a criticality hazard could potentially develop
even though only average tissile content waste was being
processed. A strategy to lessen this hazard is to mix a
small amount of neutron absorbing material into the
waste. Calculations indicated, for example, that 1.0 x
10-2 wt fraction of B203 mixed with the waste greatly
reduces the criticality hazard associated with any given
fissile inventory. This fraction of B20J uniformly mixed
with an otherwise critical concentration of plutonium in
the drying region reduces k to about 0.24, a reasonably
safe value. In fact, a similar strategy is used routinely to
prevent criticality hazards in fissile-containing solutions.
Here neutron absorption is provided by placing
borosilicate glass “raschig” rings within the solution.l
Such dramatic effects might not be obtained under all
conditions with LO x 10–2 wt fraction of B203; however,
it is clear that this simple procedure produces a highly
cost effective reduction in overall criticality hazard. In
contras~ designing boron into the gasifier structure itself
does not have a large effect on criticality. Calcu-
lationally, k of the hearth region drops from 1.00 to 0.76
when the firebrick is totally replaced with natural
elemental boron.
Calculations indicated that tissile lumps (adminis-
tratively defined as single inclusions of >200 g fissile)
can produce a rapidly developing criticality hazard. A
single 200-g lump of high density 239Pu02 can result in kof 0.21, whereas the same 239Pu02 mass spread uniform-
ly throughout the drying region produces k about 0.05.
Fissile lumps should not be processed in the SPI Facility
because they could conceivably lead to a rapidly de-
veloping criticality hazard.
B. Introduction
An SPI Facility has been proposed by the Idaho
National Engineering Laboratory (INEL) for prmxssing
their stored transuranic (TRU) wastes. Figure 1 is a
2
HEARTHMXION
SLAGREGION ‘
a PYRCLYSIS REGION
CttRAB;N
SLAG
Eiz2FIREBRICK
SCALE : t312.42cm
I
Fig. 1.
Csstcrdational geometry of the SPI, showing the yz plane at x= O. Shown are the drying, pyrolysis, and slag regions. Theslag region is composed of two smaller regions-the hearthregion and the slag tap and holding tank.
simplified drawing of the incinerator that is used for
calculational purposes. The plutonium< ontaminated
wastes enter the incinerator in the drying region, pass
through the pyrolysis region, and then enter the slag
region where the wastes are heated to a molten mass.
This incinerator model was studied for criticality haz-
ards.
These extensive Monte Carlo calculations used
LASL’S Monte Carlo code for neutron and photon
transport (MCNP).Z Specifically, MCNP was used
to calculate k of the incinerator. The k of a system is
defined as the multiplication factor of the system, The
different regions of the incinerator—drying, pyrolysis,
and slag—were studied individually and as a total
system. The k of each of the three regions was calculated
as a function of the 2J9Pu concentration. As discussed
below, the drying region of the incinerator has the lowestcritical mass of 239Pu of the three regions. For this
region, the effect of neutron poisons, namely boron in the
form B203, on k of this region was studied. Another
specialized case studied was the build-up of a layer of239pu02onthewallsof the drying and pYTolYsis regions.
The input material to the SPI Facility even after
screening could conceivably contain 200-g lumps of
239Pu02. The effect of such a lump of ‘9PuOZ in the
drying region on k of the region was calculated also.
Part of our analysis of the SPI Facility included
possible methods for monitoring the neutron activity in
the incinerator. Additional neutronic calculations mod-
eled neutron detectors surroundhig the incinerator to
determine the effect of increased moderation resulting
from the neutron detectors.
Initial parameter studies indicated the SPI Facility had
a positive temperature coet%cient; k increases as the
temperature increases. The final series of calculations for
the incinerator determined the variation of k on the
temperature of the drying region, a not insignificant
effect.
Three secondary structures associated with the SPI
Facility were also studied. These were the secondary
combustion chamber, slag tap, and holding tank; the
tundish; and the off-gas collection hopper.
All calculations were performed assuming plutonium
to be the fissile material. This is justified because the
anticipated amount of highly enriched uranium to be
processed by the SPI Facility is a small fraction of the
anticipated amount of plutonium to be processed. All
other parameters being equal, the critical mass, concen-
tration, or volume of 23gPu is less than or equal to that
for ‘3 S2’5U. Thus, results obtained for plutonium trans-
late into corresponding conservative 23SUor 233Uresults.
C. Brief Code Description
The general-purpose MCNP2 code can be used for
neutron, photon, or coupled neutron-photon transport,
which includes the capability to calculate eigenvalues of
critical systems. The code treats an arbitrary
three-dimensional configuration of materials in geometric
cells that can be specified by first- and second-degree
surfaces and some special fourth-degree surfaces (ellip-
tical tori).
Pointwise cross-section data are used. For neutrons,
all reactions given in a particular cross-section eval-
uation (such as ENDF/B-IV) are accounted for. Thermal
neutrons are described by both the free gas and S(a,~)
models. For photons, the code takes account of in-
coherent and coherent scattering, the possibility of
fluorescent emission following photoelectric absorption,
and absorption in pair production with local emission of
annihilation radiation.
Standard features that are available to improve com-
putation efficiency include geometry splitting and Rus-
sian roulette, weight cutoff with Russian roulette, cor-
related sampling, analog capture or capture by weight
reduction, exponential transformation, energy splitting,
forced collisions in designated cells, flux estimates at
point or ring detectors, deterministically transporting
pseudo-particles to designated regions, track-length esti-
mators, source biasing, and several parameter cutoffs.
Extensive summary information is provided in the
computer listing to help the user better understand the
physics and Monte Carlo simulation of his problem. The
standard, userdefined output of MCNP includes
two-way current as a function of direction across any set
of surfaces or surface segments in the problem. Flux
across any set of surfaces or surface segments is
computed. Similarly, the flux at designated detectors and
the average flux in any cell (track length per unit volume)
are standard tallies. Reactions such as fission, tritium
production, absorption, or any product of the flux times
any standard ENDF reaction cross section plus several
nonstandard ones may be obtained in any cell, at a
surface, or at a point. The heating tallies give the energy
deposition in designated cells. In addition, particles may
be flagged when they cross specified surfaces or enter
designated cells, and the contributions of these flagged
particlea to the tallies are listed separately. The user is
allowed to modify any of the standard tallies almost at
will. All quantities tallied also have their relative errors
calculated. All tallies are a function of time and energy as
defined by the user.
MCNP has the capability to calculate eigenvalues for
both sub- and supercritical systems. The calculation is
run as a series of generations of neutrons. At the end of
each generation, km is calculated for that generation as
well as averaged over a specified number of preceding
generations. The source for the first generation is usually
defined by the user at a set of spatial points in the
system. The neutrons are then started isotropically at
these points with an energy sampled from standard
fission distributions.
The cross sections used were taken from several
sources. If an ENDF/B cross-section fde was available
for the isotope of interest, it was chosen as the preferred
cross section. Some of the ENDF/B cross-section tiles
were version 111 and some were version IV. Alternative
cross-section files were taken as needed from
LLL-Howerton.
D. Material Compositions and Region Temperatures
Tables I-VI list the elemental compositions of the
materials in the different regions of the incinerator and
the secondary structures. All elemental compositions are
given as weight percentages. The elemental composition
of the material in the drying region3 is listed in Table L
The density of this material is 0.706 g/cm3. The material
in the pyrolysis region3 has a density of 1.76 g/cm3, and
its elemental composition is presented in Table II. The
slag material is common to the slag region (hearth, slag
tap, and holding tank) of the incinerator; the secondary
combustion chamber, slag tap, and holding tank; and the
tundish. The slag has a density of 2.80 g/cm3, and its
elemental composition is given in Table HI. A large
fraction of the external surface area of the slag-forming
region in the incinerator is surrounded by tirebrick.
There is also tirebrick surrounding the slag in the
secondary combustion chamber, slag tap, and holdingtank, and in the tundish. The elemental composition of
the tirebrick,3 which has a density of 2.70 g/crn3, is listed
in Table IV. The composition oft ype 316 stainless steel
on the outside of the tundish and surrounding the off-gas
collection hopper is presented in Table V. The elemental
composition of the material in the off-gas collection
hopper’ is given in Table VI. One of the compounds in
this material was a trace metal oxide, which we assumed
to be stainless steel, knowing plutonium+ontaminated
stainless steel would be part of the feed material for the
SPI Facility.
None of the above tables indicates the presence of
plutonium. The plutonium concentration, expressed as a
weight percentage, was a variable, and diiYerent amounts
were added to the drying, pyrolysis, slag, and off-gas
collection materials.
Except for the stainless steel, none of the material
compositions adds to 100%. This is because the informa-
tion provided us did not add to 100% and because the
plutonium concentration is not included in the tables.
Another parameter needed by MCNP is the tem-
perature of each material and region. At elevated tem-
peratures, neutron resonances become wider because of
Doppler broadening effects. Such an effect changes the
neutron interaction cross section. Because of relatively
high temperatures in the slag and the fwebrick, it was
deemed advisable to use appropriate temperatures. The
temperatures provided by INEL4 for the different regions
of the incinerator and for the secondary structures are
presented in Table VII.
4
TABLE I TABLE III
s
.
DRYING REGION MATERIAL
ELEMENTAL COMPOSITION*9b
Element Weight Percent
H
c
o
Na
Al
Si
s
Ca
Fe
4.99
25.10
48.54
0.57
1.62
8.15
0.80
2.02
6.86
98.65
‘Density = 0.706 g/cm3bRef. 3.
TABLE II
PYROLYSIS REGION MATERIAL
ELEMENTAL COMPOSITIONalb
Element Weight Percent
H
c
o
Na
Mg
Al
Si
K
Ca
Ti
Mn
Fe
3.53
1.41
42.30
1.41
1.06
4,23
21.20
1.41
5.29
0.03
0.46
17.60
‘Density= 1.76 g/cm3‘Ref. 3.
SLAG REGION;
SECONDARY COMBUSTION CHAMBER,
SLAG TAP, AND HOLDING TANK;
AND TUNDISH MATERIAL
ELEMENTAL COMPOSITION’lb
Element Weight Percent
c
o
Na
Mg
Al
Si
K
Ca
Mn
Fe
1.30
45.80
1.41
1.04
4.12
21.10
1.32
5.20
0.50
17.90
99.69
‘Density = 2.80 g/cm3bRef. 3.
TABLE IV
FIREBRICK ELEMENTAL COMPOSITION”b
Element Weight Percent
o
Na
Mg
Al
Si
Ca
T1
Fe
99.93
‘Density = 2.70 g/cm’bRef. 3.
49.70
0.74
0.60
20.10
26.30
0.71
1.20
0.71
100.06
TABLE V
TYPE 316 STAINLESS STEEL
ELEMENTAL COMPOSITION’
Element Weifzht Percent
Cr 18.5
Fe 68.5
Ni 13.0
100.O
.
‘Density = 8.02 ~cm3
TABLE VI
OFF-GAS COLLECTION HOPPER MATERIAL
ELEMENTAL COMPOSITION”’b
Element WeiAt Percent
c
o
Na
Mg
Al
Si
s
c1
K
Ca
Cr
Fe
Ni
1.25
37.56
15.51
0.36
3.23
21.32
1.42
13.41
0.58
1.43
0.55
3.01
0.28
99.91
“Density = 1.28 g/cm’‘Ref. 4.
E. Incinerator Geometry
Figure 1 is a view of the incinerators as modeled for
the MCNP criticality calculations. Labeled are the
drying, pyrolysis, and slag regions. The slag region
includes the hearth region and the slag tap and holding
tank. The hearth region contains the burners and is
almost totally surrounded by firebrick. The area between
the hearth and the pyrolysis regions is fdled with
TABLE VII
TEMPEIUITURE OF REGIONS
OF INCINERATOR’
TemperatureRegion [“c)
,
.
Drying 1000
Pyrolysis 1000
Slag 1482
Firebnck 1260Water Jacket 66
Tundish 1482
Secondary Combustion
Chamber 1482
Off-Gas Collection Hopper 149
‘Ref. 4.
carbonaceous char having a density of 2.1 g/cm3. There
is a 7.62-cm< 3-in.-) thick water jacket surrounding the
drying and pyrolysis regions. The slag tap and holding
tank area is surrounded by a 5.08 -cm-( 2-in.-) thick layerof firebrick and a 2.54 -cm-( l-in.-) thick water jacket. The
drying region is a cylinder having a radius of 73.66 cm
(29 in.) and a height of 256.54 cm (101 in,). The
cylindrical pyrolysis region has a radius of 73.66 cm (29
in.) and a height of 368.30 cm (145 in.). The slag tap and
holding tank area has a radius of 60.96 cm (24 in.) and a
height of 267.64 cm (105.37 in.). Other dimensions can
be obtained from the scale indicated on the figure. A
complete description of the geometry is presented in the
Appendix. ,
Criticality calculations are usually performed using a
greatly simplified geometry, replacing complicated struc-
tures with much simpler, critically conservative approx-
imations. This approach was not used in this analysis,
An accurate geometrical model was used for the entire
SPI Facility. This analysis should more accurately
predict the neutronic behavior of the facility as various
neutronic perturbations are applied to the incinerator. It
must be emphasized that this is a very complicated
geometry for MCNP calculations. Considerable
computer time is required to obtain a converged source
for k calculations. We fee[ that the 10- to 15-rein
calculations done for this study, resulting in a 2V0std dev
on k, are sufficient to allow work to proceed on the SPIFacility. However, several 1- to 2-h calculations should
be done when the plans are complete to obtain a better
“
.+
6
TABLE VIII
SPI FACILITY k CALCULATIONS
.
.
“
Case Mw ~xn k
1 7.S x 10-’ wt fi~clion ‘VII 0.49
2 1.5 x 10-1 w timabn llSPU 0.74
3 3.0 x 10-] WI fktbn ‘~Pu 0.99
4 4.65 x 10-] w tlmclbn ‘33Pu 1,16
s 3.0 X 10-’ wt frtdcm l% (6% ‘%) 0.96
Pyroly#b Region
6 6.5 x 10+ W fmction ‘Vu 0.47
7 1.3 x 10-’ W t?sdion ‘*u 0.74
8 2.2 x 10-] W fi~ctlon ‘mu 0.95
9 2.6 x 10-’ WI &actbn ‘% 1.02
SIU Radon
10 20 x 10-] WI fr~ction ‘gPu 0.s7
II 5.0 x 10-J W fmarnn ‘*Pu 0.77
12 2.0 x 10-~ w filction D*PU 0.99
SIU Tm and Holdim Tank
13 20 x 10-’ WI thctbn ‘% 0.86
14 3.0 x 10-1 w frmtbn ‘W 0.92
15 5.0 x 10-’ w frnctbn ‘% 1.00
t6 3.0 x 10-1 w fmabn “’Pui 100% boron instend of firehrick 0.92
17 2.O x 10-* WI fraction ‘~q bumwd fmbrick density 10X 1.04
18 20 x 10-1 w fimtbn ‘] Vu; 5 S/cm] HZO b wsterjnckct 0.96
Hearth ReRion
19 20 x 10-1 W fmabn ‘]’Pu 1.00
20Z. ~ ,.., ~ ~naion n~~; 100% boron inmcd of fimbrick 0.76
Critkal Cone. of ‘~11 in fich R@cm
21 3.0 x 10-] w fimtion in dryiIw rc#ion, 2.6 x 10-1 W: frwIion 0.99
in pyrolysb region, and 2.0 x 10-Z WI fraction M SIW region
converged source and to obtain a higher degree of
confidence in the results.
F. Results
The results of the MCNP k calculations are presented
in Table VIII. Unless noted otherwise, all calculations
were made for 239Pu. All k values have a 2V0 std dev
associated with them. Cases 1-5 give the variation of kwith plutonium concentration in the drying region of the
incinerator. The calculation in Case 5 was made using
weapons-grade plutonium (94% 23gPu, 6% 240Pu) in the
drying region. The variation of k with the plutonium
concentration in the pyrolysis region is given in Cases
6-9. Studies of the slag region, slag tap and holding tank,
and hearth region are summarized in Cases 10-20,
Included in these slag region studies are the effects of
case timg R@on whh 3.o x Itr~ w fraction ZI~pU k
22 1.0 x 10-1 w frmtio” BIO, in drying ~~ott 0.77
23 1.0 x 10-1 WI fraction B,O, In drying region 0.2A
24 5 S/cm] H,O in wstcr jacket 1.01
25 10096 boron in water jickd 0.96
26 incrawd HZO in drying rtgion from 2.5 W% to 37.5 w% 0.9s
l-cm hycr WU02 “mMyiq and Py?olysk Regions
0 (WUO.1 = 11.5 ticm]
27 no other ‘~Pu in incimmtor 0.97
2s 100% bmon in watt? Jeckcl; no other ‘]% 0.s029 5 #cm] HIO in water jcckct: no other l]JPu 1.02
2013.& 1.61 an.RuIius, Spherical Lump “PuO, M Drying Region
p (WUO) = 1 M #cm’
30 3.0 x 10-] WI fmctbn “*Pu in drying region 0.9s
31 no olh Wu in incimmtw 0.21
3.0 x 10-1 WI fraction 1]% in Drying Reglo”
32 500”C drying tcmpcmtwe, all other tctnpaaturcs ●t ~dard 0.88
Vducs
33 750”C 0.96
34 loa3”c 1.01
3s 1170”C I.00
36 Isoo”c 1.03
06-Gs Colktion HOPFCr
37 5.0 x 1O-t W fraction ‘gPu 1.02
Tundish
3s 4.0 x 10-] M fiacdon ‘*Pu 0.97
secondary Combustion Chamber, SIM Tw, and Holding Tank
39 20 x 10-1 W fraction ‘% 0.S6
40 4.0 x 10-Y W baction ‘~Pu 0.9741 5.0 x !O-’ wi fraction ‘Wu 1.03
changing the external neutron moderators. Case 21 had
critical concentrations of 23gPu every where.
Some calculations were made to study the effects of
adding neutron absorbing material to the feed mixture on
the critical concentrations of 23gPu. These studies were
done only for the most critically sensitive region, the
drying region, and are summarized in Cases 22 and 23.
The effect of external neutron moderators on the neu-
tronics of the drying region is listed in Cases 24 and 25.
Case 26 indicates what happens to k when the amount of
water in the drying region is increased from 25 to 37.5
wt’%o, keeping the weight fraction of 239Pu constant.
A potential nonstandard operating situation is the
build-up of a layer of plutonium on the inside surface of
the drying and pyrolysis regions. A very conservative
estimate was made assuming this layer to be 239PuOZ of
density 11.5 g/cm3. The results of the calculations for a
layer of 239Pu02 are presented in Cases 27-29.
7
The feed material for the SPI Facility could contain
lumps of plutonium having a mass up to 200 g. The
‘3gPu02 (p = 11.5 g/cm3) on keffects of a 200-g lump of
of the drying region are presented in Cases 30 and 31.
Calculations indicated that k is sensitive to the tem-
perature of the region. Cases 32-36 summarize the
variation of k with the temperature of the drying region.
The last five entries in Table VIII, Cases 37-41, are
the results of calculations on the off-gas collection
hopper (Case 37), the tuncfish (Case 38), and the
secondary combustion chamber, slag tap, and holding
tank (Cases 39-4 1).
G. Discussion of Incinerator Results
Figure 2 shows the variation of k with 23gPu concen-
tration in the drying region of the incinerator. The drying
region has a critical 23gPu concentration of 3.0 x 10-3 wt
fraction. This is equivalent to a mass of 9.3 kg 239Pu.
When the plutonium concentration decreases by a factor
of 2, k of the system drops to 0.74. Another decrease by
a factor of 2 in the plutonium concentration decreases k
to 0.49.
Because the drying region has the lowest critical 23gPu
mass, this region was used to study matrix effects. The
SPI Facility wiU be processing stored wastes that are
contaminated mainly with weapons-grade plutonium.
Consequently, a calculation was made to determine the
effect of 6°A 2@Pu. The result of this calculation is shown
in Fig. 2, detailing the variation of k in the drying region
with plutonium concentration. This calculation assumed
a critical concentration of plutonium, 3.0 x 10-3 W
fraction. Weapons-grade plutonium lowers k of the
drying region by about 3Y0. Plutonium-240 has a fission
threshold of about 1 MeV; thermal neutrons can fission
ic-___l0 la,lo-s 1,0,10-, 3.0.10-, 4,0,10-M.to+
WHGNT FISACTSON ‘a% USDRVINQ REGION
Fig. 2.Variation of k for the ckying region with the weight fraction ofIlgpu ~ t~ ds.ying region. ● = ‘9Pu; A = 94% 23Tw 6%
MOpu. me ~~e of t~ data poims includes the 2% std dev on
the cnkulations. The curve merely serves to guide the eye.
239Pu. The drying region is 5 wt’Yo hydrogen or 45’%0
atom fraction hydrogen. Thus the drying region is
heavily moderated, and the neutrons rapidly lose energy.
Neutrons are not able to fission 24Pu and there is less
23gPu in the region. Consequently, k of the region
decreases.
Another matrix study increased the amount of water
in the drying region. One class of stored wastes is sludge
drums containing a large amount of water. The effect of
additional water on k of the drying region needed to be
quantified. The drying region is normally 25 wt’%owater.
The amount of water in the drying region was arbitrarily
increased by 50?40to a total value of 37.5 wtYo. The k of
the drying region under this situation is 0.98. The
increase in water did not substantially change k of the
region. (The difference is within the standard deviation
on the MCN P calculations.) The initial hydrogen densit y
in the drying region thermalizes most of the fast fission
neutrons. Adding hydrogen does not produce any sub-
stantial additional thermalization.
The effect of adding neutron absorbers on k of the
drying region was another matrix study. Varying
amounts of BZ03 (natural isotopic composition) were
added to the material in the drying region. For this series
of calculations, the 23gPu concentration in the drying
region was held f~ed at the critical concentration, 3.0
x 10-3 wt fraction. Figure 3 summarizes the effect of
adding B203 to the material in the drying region. A 1.0
x 10-3 wt fraction of B203 reduces k from 0.99 to 0.77.
A 1.0 x 10-2 wt fraction of B20J xeduces k to 0,24. A
small amount of natural boron added to the feed material
appears to be a very effective technique for controlling
the criticality of the SPI Facility. When a LO x 10-2 wt
1,W
0,s0 -
x 0.10 -
0.08 -
0,01 t , , , 9,1 ,1 J
s.o.lo-~ 6.0s 10-* 1.0.90-s b.o.l o-~ t O.,0-* m.$o-a 1.0’% 0-8
WEIGHT FRACTION ml OS IN DRYING REGION
Fig. 3.Effect of adding BZOJto the drying regionon k of the dryingregion. The size of the data points includes the 2% std dev onthe calculations. The curve merely serves to guide the eye.
.
8
fraction of BZOJ is in the feed mixture, the neutron
spectrum that would be incident on a detector is
somewhat harder, and there are about 15% fewer
neutrons emerging from the drying region. For heavily
moderated 3He tubes, the boron will probably cause
about a 159’0change in the expected signal.
The feed material for the drying region could contain
lumps of plutonium. The maximum permissible loading
of plutonium in a 208-t’ (55-gal.) drum is 200 g. A
critically conservative approach for these studies con-
sidered a 200-g spherical lump of 239PuOZ with the
theoretical density of 11.5 g/cm3. Such a sphere of
239PuOZ has a radius of 1.61 cm. The size of this sphere
is very small compared to the volume of the drying
region of the SPI Facility, For MCNP to adequate-
ly sample this small volume of ‘9PuOz, special im-
portance sampling techniques were used. Two calcu-
lations were performed on the lump of 239PuOV One
calculation assumed the lump was in the drying region;
the region had a critical 239Pu concentration of 3.0
x 10-3 wt fraction. Without the lump, k of the region is
0.99; with the 200-g lump of 239pu02, k of the region’s
0.98. No effect outside the 2?40 std dev on the calcu-
lations was observed. The other calculation on the 200-g
lump of zjgpu02 ~sumed there was no other 239Pu in the
drying region. This situation has k = 0,21. Additional
criticality calculations must be performed if lumps of
plutonium having a mass larger than 200 g are to be
allowed in the feed material.
The use of an external neutron absorber to control the
criticality was studied by replacing the water in the water
jacket surrounding the drying region with natural boron.
A critical 239Pu concentration of 3.0 x 10-3 wt fraction
was in the region. Using external means to control
criticality is very ineffective. The boron reduced k of the
system from 0.99 to 0.96, not a statistically significant
change.
The effect external neutron detectors would have on
the neutronics of the SPI Facility was studied also. The
assumption was made that thermal neutron detectors
would be used to monitor criticality. Thus there would be
slabs of polyethylene or other efficient neutron mod-
erators surrounding the incinerator. These external mod-
erators were modeled by increasing the density of the
water in the water jackets. No attempt was made to take
into account that neutrons would be absorbed in the
detectors. This is a conservative approach. Increasing
the density of water in the water jacket surrounding the
drying region changed k for this region from 0.99 to
1.01. This change is not statistically significant.
I I I1.0
0.8
0.6x
/0.4
0.2
I/ 1o.oo~
2.ox1o-3 3. OX1O-3
WEIGHT FRACTION 2391%1 IN PYROLYSIS REGION
Pig. 4.Variation of k for the pyrolysis region with the weight fractionof Ugpu in the pyr~ysis region. The size of the data points
inoludes the 2% std dev on the calculations. The curve merelyserves to guide the eye.
The variation of k with 239Pu concentration in the
pyrolysis region is shown in Fig. 4. The pyrolysis region
has a critical 239Pu concentration of 2.6 x 10–3 w
fraction. This is equivalent to a mass of 28.7 kg 239Pu. A
decrease by a factor of 2 in the ‘9Pu concentration
reduces k to 0.74. Another decrease by a factor of 2 in
the 23%% concentration further reduces k to 0.47. This is
the same relative behavior as observed in the drying
region.
A conservative approximation was made of the effect
of plutonium build-up on the inside surfaces of the drying
and pyrolysis regions on k of the system. This is an
abnormal operating condition. The layer of plutonium
was 239Pu02 (theoretical density p = 11.5 g/cm3). Any
actual layer of 239Pu02 on the inside surfaces would
probably have a density closer to 2.0 g/cm3. The plating
of plutonium on the inside surfaces of the incinerator
does not create a criticality hazard per se. Our calcu-
lations indicate that a large mass of plutonium (1 O’s of
kg) plated on the inside surfaces is required to create a
signitlcant criticality hazard. Build-ups of this magnitude
would be detected by the neutron detectors surrounding
the incinerator long before any hazard develops. The real
criticality hazard is the potential of this layer of
plutonium breaking up and falling down into the in-
cinerator. Nine kilograms of plutonium uniformly dis-
tributed in the drying region is critical. Using the neutron
detectors discussed below, 9 kg can be detected on the
walls of the incinerator. Assuming a safe operating
condition of k ~ 0.20, 800 g can be plated on the walls,
and a criticality hazard will not develop if that quantity
falls into the incinerator, even if all the material coalesces
into one lump.
9
o,.o~0.s110-2 l.o,lo-~ l.sxlo-a 2.0s10-2
WEIGHT FRACTION 239Pu IN SLAG REGION
Fig. 5.Dependence of k on the weight fraction of 2’9Pu in the slagregion The sir.c of the data points includes the 2% std dev onthe calculations. The curve merely serves to guide the eye. . =Zlgpuin the tot~ slag region. For the critical concentration 2.0
x 10-2 w fraction, k was calculated for ‘]9Pu only in thehearth region (0) and fm “% ordy in the hearth region withthe tirek”ck surrounding the hearth replaced with 100%natural boron (A).
Figure 5 shows the variation of k with 239Pu concen-
tration in the slag region. The slag region has a criticalZJ9pu concentration of 2.o x 10–2 wt fraction. This is
equivalent to a mass of271 kg 239Pu. For the slag region,
a decrease by a factor of 4 in the 239Pu concentration
redu~es k to 0.77. A further decrease by a factor of 2.5 in
the 239Pu concentration reduces k to 0.57. Thus the slag
region is not as sensitive to changes in 23% concentra-
tion as are the drying and pyrolysis regions. However, it
is ex~cted that the 1.0 x 10-2 w fraction of B20J
placed in the slag region will make a similar reduction in
k as it did in the drying region. For these calculations of
k, total blockage to the level of the burners was assumed.
The critical mass of the slag region is so large mainly
because of the slag tap and holding tank. Blueprints of
the slag tap and holding tank had a dimension labeled
“as required:’ which was assumed to be 85 cm.
Several studies for the slag region were made because
sampling problems were observed when the entire region
was treated as a whole. Subsequent calculations were
made only on the hearth region and only on the slag tap
and holding tank area. (See Fig. i for a definition of
these regions.) Figure 5 compares the results for the
hearth region with those for the total slag region. The
comparison is for a single 239Pu concentration, 2.0x 10-2
wt fraction. This weight fraction is critical for the total
slag region, and there is no difference in k for the hearth
region and the total slag region. The hearth region has a
critical mass of about 80 kg 239Pu. When the tirebrick is
x 0.6v ● Is*p” ,“ ,~,~~ .$LAO REC*”
8 a3mPu ONLY M SLAO TAP, p(lb6bfkh):27glcmS
0.4 A *s*p.ONLY IN SLAO TAP. plnzo):swcma1
0.2u ~23*Pu ONL” 1“ SLM TAP
o..~o o.sslo-~ 1.0,10-2 1.5,10-2 2.0.10-1
WE,GHT FRACTION 239Pu IN SLAG REGION
Fig. 6.
Dependence of k on the weight fraction of *3% in the slagregion. The six of the data points inchsdes the 2% std dev onthe calculations. The curve merely serves to guide the eye. . =z19pu in the to~ slag region. Additional calcUiatiOrrS Were
made for the critical concentration 2.0 x 10-2 wt fraction. The4 gives k for ZWPU Omy in the slag tap. The A gives k when
th=e is ord y239PU ~ the ~ag tap, md the density of water in
the water jacket surrounding the slag tap has been increasedby a factor of 5. The ■ gives k for 239Puonly in the slag tapwhen the density of the fuebrick surrounding the slag tap hasbeen increased by a factor of 10.
replaced totally with natural boron, k of the hearth
region drops significantly to 0.76.
The comparison between the total slag region and the
slag tap and holding tank only is shown in Fig. 6. Whena 2.() x 10-2 Wt fraction of 239Pu is present only in the
slag tap and holding tank, k of the incinerator is 0.86.
For this same concentration in the total slag region,
k = 0.99. Thus k of the total slag region is effectively k
of the hearth region which is the most reactive region. To
a certain extent, the hearth region is neutronically
isolated from the slag tap and holding tank. This is
partiaUy due to a feature of the calculational model. The
blueprints of the slag tap region had a dimension “as
required.” As this dimension is increased, the two regions
of the slag become more isolated.
The k of the slag tap and holding tank is lower
because of the absence of adequate neutron moderators
surrounding the slag tap and holding tank. They are
surrounded by only 5.08 -cm-(2-in.-) thick fuebrick and a
2.54+ m-( l-in.-) water jacket. When the density of the
water in the water jacket surrounding the slag tap and
holdipg tank is artificially increased fivefold, k increases
to 0.96. This effect increases the neutron moderation
near the surface of the slag tap and holding tank and
increases the number of fissions near the surface.
Artificially increasing the density of the water also
simulates the effect of a thermal neutron detector,
.
.
10
heavily moderated with water or polyethylene, used to
monitor activity in the slag tap and holding tank region.
The effect of external neutron moderators was estab
lished also by artificially increasing the density of the
firebrick surrounding the slag tap and holding tank. A
tenfold inciease in the tirebrick density increases k to
1.04. The effect of external moderators around the slag
tap and holding tank is so noticeable because the slag
does not have any efficient neutron moderators in it. The
interior of the slag tap and holding tank appears to be an
infinite medium, but the outer few centimeters of the slag
1.03 * 0.02 for T = 1500”C. There is a noticeable rise ink as the temperature rises from 500 to 1000”C. For
temperatures above 1000° C, k is much less sensitive to
changes in temperature. As the temperature rises from
1000 to 1500”C, k varies within the 2?40 std dev on the
calculations. Because the drying region operates at
1000”C, any increase in temperature should produce
only small fluctuations in k. Any decrease in operating
temperature from 10OO°C decreases k. These calcu-
lations emphasize the need to use correct temperatures
for the different regions of the incinerator and for the
tap and holding tank are easily influenced by external cross-section evaluation because of the considerable
moderators. Doppler broadening effects on neutron resonances.
After the critical ‘9Pu concentration of each region
was determined, a calculation was made on the entire
incinerator with each region having a critical concentra-tion of 239Pu in h. l’he Ic of the entire system was 0.99.
Within the 2L% std dev on the calculations, there is no
difference between calculating k for each region separate-
ly and calculating k for the entire incinerator. There is
little neutron transfer between the different regions.
Neutronically, each region appears to be independent.
We are thus justified in treating each region separately.
Figure 7 summarizes the temperature effect on the
incinerator, specifically the drying region. The drying
region has an operating temperature of 1000° C. For the
calculations summarized here, the temperature of this
region was varied from 500 to 1500° C. The tem-
peratures of all other regions were held fixed at the
values listed in Table VII. For all temperature calcu-
lations, the drying region had a 2’9Pu concentration of
3.0 x 10-3 wt fraction, a critical concentration when thetemperature is 1000”C. The k of the drying region varies
from a low of 0.88 + 0.02 for T = 500”C to a high of
I 1 I t I 11.0s
1
1.00
0 8s
x 1
H. Secondary Combustion Chamber, Slag Tap, and
Holding Tank L-
The calculational model of the secondary combustion
chamber, slag tap, and holding tankc is shown in Fig. 8.
The secondary combustion chamber, slag tap, and
holding tank has cylindrical symmetry. The lower region
is surrounded by a 5.08-cm-(2-in.-) thick fwebrick, which
is surrounded by a 2.54-cm-( 1-in.-) thick water jacket.
The compositions of the slag material and the tirebrick
were presented in Tables III and IV, respectively. Figure
9 shows the placement of materials in the secondary
combustion chamber. From Table VLII, the calculated
critical 23gPu weight fraction in the secondary combus-
tion chamber, slag tap, and holding tank is between
4.0 x 10-2 and 5.0 x 10-2. This is equivalent.to a mass
between 450 and 560 kg 2~9Pu, This 239Pu weight
fraction is to be compared with a 2.0 x 10-2 critical wtfraction of 239pu in the slag region. The secondary
combustion chamber, slag tap, and holding tank unit
dces not have as much moderating material around it as
does the slag (hearth) region. Consequently, the critical
concentration of 239pu in the ~condary combustion
chamber, slag tap, and holdlng tank is greater than that
in the slag region.
L Tundish
0 no
00 I I I 1 I I 10 2s0 Soo 7S0 1000 1250 1500 The tundish4 is a parallelepipeds with dimensions
TEMPERATURE OF ORVING REGION (“Cl 426.72 cm (168 in.) x 365.76 cm (144 in.) x 213.36 cm
Fig. 7. (84 in.) high. The bottom and sides are surrounded by a
Variation of k with the temperature of the drying region, which 60.96-cm-(24-in.-) thick firebrick. Surrounding the fwe-
contained a 3.0 x 10-] wt fraction of’2’QPu. The curve merely brick is a 5.08 -cm-(2-in.-) thick type 316 stainless steelserves to guide the eye.
--l L40.64 cm
al~ 60.96 cm66.04cm68.5ecm
Fig. 8.Calculationcd geometry of the secondary combustionchamber, slag tap, and holding tank. Appropriate dimensionsof important regions are given.
FIREERICKA
/’
/—VACUUM
\
~ SLAG
—FIREBRICK
—WAIER JACKET
— VACUUM
Fig. 9.Matcritd regions In the secondary combustion chamber, slagtap, and holding tank.
jacket. The material in the tundish is identical to that in
the slag region of the incinerator, which was presented in
Table III. Initially, the tundish would be expected to
behave neutronically similar to the slag region and the
secondary combustion chamber, slag tap, and holding
tank. However, the criticaI 239Pu concentration in the
tundish is an order of magnitude less (see Table VIII).
The critical wt fraction in the tundish is 4.0 x 10-3. This
‘9Pu. The slag does not have anyis equivalent to 370 kg
efficient neutron moderators in it. The interior of the
tundish appears infinite to the neutrons and thus is not
affected by surrounding material. This is not true of the
slag material near the outside edges of the tundish. The
tundish is surrounded on five sides by fuebrick and hasmore moderating material around it than does the slag .
region of the incinerator or the secondary combustion
chamber. The firebrick surrounding the tundish is thick
enough to moderate and reflect neutrons into the tundish.
These reflected neutrons are of the right energy to fission
the 239Pu near the edges of the tundish, enhancing the
fission rate and lowering the critical 239Pu concentration.
J. Off-Gas Collection Hopper
The off-gas collection hopper’ is a 152.40-cm-(60-in.-)
diam, 182.88 -cm-(72-in.-) high cylinder, The walls of the
collection hopper are 0.48-em-(0. 19-in.-) thick type 316
stainless steel. The composition of material in the
collection hopper was presented in Table VI. The critical
mass of 239Pu in the collection hopper is calculated to be
50% by weight (see Table VIII), This is equivalent to a239Pu mass of 2140 kg, which initially seemed too large a
mass. However, the material in the collection hopper
does not contain any neutron moderators, and the
collection hopper is not surrounded by neutron mod-
erators. Several secondary calculations were made,
artificially adding hydrogen to the composition. As the
amount of hydrogen was increased, the critical mass of239Pu decreased. When the amount of hydrogen ap
preached that in the drying or pyrolysis regions, a
plutonium concentration similar to that in the drying or
pyrolysis regions was calculated.
K. Conclusions from Monte Carlo Studies
If only properly screened input material is processed,
there does not appear to be a serious criticality hazard
for the SPI Facility. The region having the largest
criticality hazard is the drying region, where the critical
mass of 239Pu is 9.3 kg. The next most hazardous critical
region would be the pyrolysis region, which has a critical
mass of about 30 kg ‘39Pu. All other regions of the
incinerator and seeondary structures have critical masses‘9Pu. Forty-five 208< (55-gal.)greater than 200 kg
drums, each containing 200 g of plutonium, could be put
into the SPI Facility at one time before causing a critical
situation. This assumes that no build-up occurs.
The average plutonium concentration expected in the
SPI Facility is 1g plutonium per 50 kg waste. This is a wt
,
.
.
.
12
fraction of 2.0 x 10-s. There is no criticality problem in
any region of the SPI Facility for this average waste. The
drying region, which has the lowest critical mass, has a
critical weight fraction that is one hundred times larger
fhan that of the average waste. Other regions of the SPI
Facility have critical weight fractions two-tcthree orders
of magnitude greater than that of the avefage waste..However, one simple operation would greatly reduce
any conceivable criticality hazard. A small amount of
BZOJ ( 1.0 x 10-2 w fraction) added to the feed mixture in
the drying region reduces k from 0.99 to 0.24,
The addition of water to the feed material of the
incinerator does not produce any statistically significant
change in k of the drying region. Thermal neutron
counters are expected to have minimal effects on the
neutronics of the incinerator. Likewise, the use of
external neutron absorbers to control k of the incinerator
is not an effective technique, Such a technique has about
a 20% maximum effect on k. To get this effect, either the
water in the water jacket or the firebrick must be
replaced totally with boron. Such changes a~ not
possible because the water is needed for its cooling
action, and the firebrick is needed for its
thermal-insulating properties.
Lumps of plutonium, taken as 200 g or less of
~9PuOz, will not pose a criticality problem. It takes a
large amount of plutonium build-up on the insi& sur-
faces of the drying and pyrolysis regions to create a
criticality hazard. However, the real criticality hazard
with plutonium build-up develops if the layer of
plutonium falls off the wall into the incinerator. Any
plutonium build-up can be monitored by detectors sur-
rounding the SPI Facility and alarms sounded long
before any hazard develops.
The incinerator itself, excluding the tundish; secon-.dary combustion chamber, slag tap, and holding tank;
and off-gas collection hopper, has a positive temperature
coefficient. An increase in the temperature of a region
increases k of the region. This increase in k is most
noticeable as the temperature in the drying region
increases from 500 to 1000”C. Above 1000°C, k
appears to level off with increasing tem~rature. The
temperatures of the plutonium-containing regions are
between 1000 and 1500°C, Thus changes in temperature
should not produce large changes in k. For the drying
region, an increase from 1000 to 1500° C changes k by
at most 2%.
II. DIRECT CRITICALITY MONITORING
A. Section Summary
We define direct criticality monitoring as the measure-
ment of radiation directly associated with the fissile
material producing the criticality hazard. This is dis-
tinguished from indirect criticality monitoring, which
uses the differences between input and output assays to
infer residual fissile mass in the facility.
We examined both passive and active techniques for
direct monitoring of system criticality, including total
count rate, Feynman Variance, and Rossi-alpha. We
identitie~ based on the very sofl neutron spectrum
emerging from the drying and pyrolysis regions, a
particularly appropriate neutron detector design that can
be used for both passive and active measurements. We
calculated responses for these detection systems for
efficiencies typical of the SPI Facility criticality monitor-
ing geometry and demonstrated with measurements on
several plutonium multiplying systems the feasibility of
this type of monitoring system. If the criticality hazard is
judged to be suiliciently serious in other regions of the
SPI Facility, a similar neutron monitoring system can be
easily installed and operated in those regions.
A passive neutron system should be adequate for
direct criticality monitoring. The passive neutron data
would be analyzed using both the total neutron count
rate and a time-correlated analysis. The former provides
a sensitive, very timely monitor of the system. A
time-correlated analysis is used to minimize the in-
terference from neutrons from matrix generated (a,n)
reactions. An upper limit alarm level can be set using the
total count rate while the time-correlated analysis is used
to determine, with a longer time constan~ the actual
system multiplication.
A compact pulsed neutron generator can be in-
corporated into the system. This feature allows both the
system multiplication and the tissile material build-up to
be determined. It is clearly a more complicated system
than the passive system.
B. Passive Neutron System
The MCNP calculations of the SPI Facility in Sec. I
included neutron spectral output in the regions external
to the drying and pyrolysis regions. These calculations
indicate a very soft neutron spectrum, produced primari-
ly by interactions with hydrogen within these regions and
in the water jacket surrounding them. Depending on the
location, 1/3 to 1/2 of aIl neutrons in the potential
detector region are near thermal energies. For this
situation, an appropriate neutron detector is shown in
Fig. 10. This design consists of an “open-faced sand-
wich” of ‘He proportional counters and polyethylene.
The bare portion of the 3He proportional counters faces
the source and has a calculated efficiency’ of about 95%
for incident thermaIized neutrons. The efllciency falls off
rapidly with energy; the energy-averaged efficiency for
the calculated drying/pyrolysis region spectrum is about
40-50V0.
Because the neutron count rate is proportional to
detector solid angle, the largest possible solid angle
should be intercepted in the drying/pyrolysis regions. A
2- 10% effective solid angle is possible, resulting in a 47t
detection efficiency in the range of 1-5’?40.
The total passive neutron count rate can be used as a
criticality monitor. Historically,aq9 just such a system
was used to monitor the state-of-the-system criticality in
fast critical assembly research and development work.
The typical SPI Facility waste matrix could produce
Fig. 10.Typical set of ‘He/polyethyleneopen-facedsandwichneutrondetectors. This configuration k optimized for a highly mod-erated incident neutron sFcctrunL
(a,n) reactions. If a particular load of waste were to
contain significant amounts OC for example, Be, B, or F
(a very credible possibility ’~, then any TRU present can
induce (ajn) reactions. Table IX (from Ref. 11) gives the
measured neutron yields for typical TRU alpha particles
incident on various materials with low atomic numbers.
A batch of waste containing modest amounts of
plutonium combined with significant amounts of Be, B,
or F could give a total neutron count rate that would
appear to signal the presence of a significant criticality
hazard. This would constitute a false alarm for a
monitoring system based solely on the total neutron
count rate.
Because the very nature of system multiplication
guarantees a large number of time-correlated neutrons,
one solution to the false alarm problem produced by
(a,n) neutrons is time-correlation measurements. Many
techniques have been devised to accomplish this. lZI*3Of
particular applicability, however, is the Feynman or
Reduced Variance technique. 14S]SThis technique has
been developed at LASL in connection with a Nuclear
Emergency Search Team (NEST) requirement for
portable system multiplication measurement equipment.
A prototype module of such a system is shown in Fig.
11.
From an analysis of the basic physics of a multiplying
system, one can show that *4
K(tk -1)Y = const. —-— s
#(1)
TABLE IX
THICK TARGET, NEUTRON YIELDS
FOR 5.3-MeV ALPHA PARTICLES’
Tareet Material Observed Neutron Yield/ 106 a
.
Li
02Na
Mg
Al
Si
c
F,
B
Be
2.60
0.07
1.50
1.40
0.74
0.16
0.11
12.00
24.00
80.00
●
✎
“Refl 11.
14
Fig. 11.Front panel view of microprocessor-based Feynman Variance
data collection/analysis module.
where M is the system multiplication, ~ is the average
number of prompt neutrons emitted in a neutron-induced
fission of the t%sile material, and Y is a quantity
generally defined as the “variance-to-mean ratio.”
Experimentally,
-2C-C2 2iiy.— +—- 1,
i T(2)
where ~ is the experimentally measured average
counts-per-random gate, T is the length of the randomly
generated gates, ~z is the second moment of the
observed distribution, and r is a system deadtime. The
system deadtime is measured experimentally by placing a
time-uncorrelated source, such as an (cqn) source, near
the detection system. For this source, Y = O and r may
be obtained.
In the experimental situation of a mixed (a$n) and
f~sion source, the foUowing quantity is useful.
(3)
A comprehensive study (theoretical and experimental) of
the general problem of mixed (a,n), spontaneous fission
and system multiplication has been completed, and an
initial report is being prepared. 16
As a demonstration of the effectiveness of the Feyn-
man Variance technique to determine system muki-
plication, we completed a series of measurements on
some plutonium samples ranging in mass from about 0.5
to 4.0 kg. The cylindrical counting geometry was similar
to that expected for an implemented neutron detection
system in the SPI Facility around either the pyrolysis or
drying regions. However, no attempt was made to mock
up matrix effects for these measurements. That level of
effort was beyond the scope of this study. These
measurements serve to demonstrate the feasibility of the
technique for plutonium masses of a size (as indicated by
our MCNP calculations) that could be expected to result
in significant system multiplication. This method is used
routinely in the LASL NEST effort.
Figure 12 shows a plot of the basic experimental data
with the quantity 1 + Y plotted as a function of
plutonium mass. This results in a scatter plot with a
I 1 [ f II
1.0 -
1.7 -●“
. ● ..*,.”
1.6 -●.@
●.*
● 1.5 -
+. ●.4*”*’
- 1.4 -.“
●’*” .,.”
1.s - ●..*”4. . .
..“
1.2 - ●.4 ● ●
,.~,.
1.1 - ●**e “
.“OO t I 1 I1 2 3 4
Pu MASS (Kg)
Fig. 12.Feynman Variance data (1 + Y) obtained for a set ofmultiplying plutonium samples and plotted as a function ofplutonium mass. Detection system used cylindrical geometry.
15
generally increasing trend as a function of increasing
plutonium mass. The reason for the large scatter is that
plutonium mass alone is not a good indicator of system
multiplication. Some of these samples were spherical,
some were cylindrical, and the sample density and ‘40Pu
isotopic content varied greatly.
When the same data are plotted dfierently (see Fig.
13), a very smooth curve results. Here 1 + Y is plotted as
a function of the quantity Q ~ ‘OPu mass. Dividing bythe ~’Jpu ma% which was known for all the .SSmples,
normalizes all samples according to the “driving source”
present— 240pU sPon~neous fission. If the s~Ples Were
nonmultiplying, all would yield the same 1 + Y value.
The actual system multiplication for each of the samples
used was not determined independently. However, a
MCNP calculation of the largest sample was made and
indicated M ~ 3.0 (or k s 0.67). This corresponds to the
largest 1 + Y value measured, about 1.9. Using this
value to establish an approximate multiplication for the
lowest 1 + Y value measured leads to an estimate of M ~
1.2 (k ~ 0.17).
In Fig. 14, we show the calculated Feynman Variance
quantity Q px fission as a function of system muki-
plication M for the anticipated range of 4X detection
efficiencies. We would estimate the technique to besensitive to k of 0.05-0.10 (M = 1.05-1.11).
To determine 1 + Y, we generally acquire data for 10c
random (but back-to-back) gates. The gate length must
be matched to the detector/SPI Facility dieaway time.
For the SPI Facility monitoring system, we estimate that
T ~ 1 ms is appropriate. This indicates a 1000-s count
time is needed to acquire a typical set of data. However,
II I I
Y- 1
I 1 1 I 1 1
.0s .10 .15 .20
0 + 240Pu MASS (Kg)
Fig. 13.
Same Feynman Variance data (1 + Y) as shown in Fig.now plotted as a function of Q divi&d by 2aPu mass.
16
12
for a fixed installation and use as a criticality monitor, a
technique of time averaging would be appropriate. We
have used this type of analysis successfully in other field
instrumentation but not in the Feynman work. Here one
could use a 1000-s analysis period, for instance, but
update it more frequently. For example, a 1 + Y could be
calculated every 5 or 10s, with exclusion of the “oldest”
5 or 10 s of data and inclusion of the “newest” 5 or 10s
of data. If a significant change in 1 + Y occurs in the 5-
or 10-s period, this would be reflected in the latest 1 + Y
value because a few highly correlated events (neutron
chains) would be weighted very strongly in the 1 + Y
calculation. On the other hand, by retaining a relatively
long total count time, much better counting statistics are
obtained. This avoids the frequent false alarm problem
that is always associated with poor counting statistics.
Another approach wpuld be to set up a combined
“AND” alarm system in which both a high 1 + Y value
and a high total count rate are required before an alarm
is sounded. This AND alarm approach has been used
successfully at LASL. One can be sure that the most
severe operational problem associated with a criticality
monitor (or any other monitor) will be the false alarm
problem. It is beyond the scope of this study to
recommend a more specific solution to this false alarm
problem. We believe, however, that the basic sensitivity
MULTIPLICATION M
Fig. 14.
Calculated Feynman Variance quantity (Q per tlssion) as afunction of system multiplication. The 4Xdetection eftlciencies(0.01, 0.02 0.03) are in the expected range for the SPIcriticality monitoring system.
is well in hand with the 3He/polyethylene open-faced
sandwich detectors and that some combination of total
count rate and time-correlated (Feynman Variance)
analysis will be sufficient. The analysis procedures are
available at LASL and appropriate hardware exists.
These are discussed in the references cited above. iz-]b
.
C. Active Neutron System
If the passive neutron criticality monitor cannot
provide a rapid enough response time, a pulsed active
system might be required. Pulsed neutron sources based
on the D + T reaction are commercially avai1able17 and
have been used for various safeguards and activation
analysis purposes. When the neutron correlation meth-
ods are reduced to their essence (Rossi-alpha, Feynman
Variance), one finds that the pulsed neutron theory is
identical to that used for a passive analysis. One
measures various features of neutron multiplying chains,
and it matters little if the initiating event occurs sponta-
neously or is induced.
The Rossi-alpha measurement generally involves
measuring the prompt dieaway time of a multiplying
assembly. The 3He/polyethylene open-faced sandwich
neutron detectors discussed above are capable of
monitoring the prompt dieaway time and could be used
in an active pulsed mode. Fast recovery electronics are
available (currently in use at LASL and elsewhere) for
these detectors so that they can be operated in the
presence of an interrogating pulsed source. One needs to
keep open the active neutron option. A pulsed neutron
generator of more than adequate neutron output ( 1 x 106
n/pulse, 100 pps pulse rate) is currently available. ” This
system is quite small. It was designed to fit down small
diameter boreholes for the oil well and uranium pros-
pecting industries. Thus, it could be installed at the SPI
Facility around the pyrolysis or drying region without
occupying much space. Indeed, it is only slightly larger
than a single 3He proportional counter.
HI. DISTINGUISHING BETWEEN FISSILE AND
FERTILE MATERIALS IN NUCLEAR WASTE AT
ALL STAGES
A. Section Summary
Fissile material is best distinguished from fertile mate-
rial by using assay techniques that are fissile specific. In
this regard, techniques based on thermal neutron induced
fission provide the solution. Such techniques specifically235u 239pu, 241pu, ad other hk isotopes!
measure ,
while giving no response for the common fertile isotopes,
such as 23EU. Techniques based on thermal neutronfission and directed specifically to measurements in
typical waste matrices 17-*9are currently under develop-
ment. One or more of these active neutron techniques
should sutlice to determine the fissile material content in
the variety of wastes the SPI Facility is likely to
encounter. These techniques can probably be developed
to handle any particularly diflicuit matrix (such as
sludges or concrete) in the leadtime available before the
SPI Facility is scheduled to be operational.
As an adjunct to thermal neutron interrogation tech-
niques, passive neutron coincidence methods can be used
to quantify plutonium content in waste of reasonably
well-characterized plutonium isotopic content, such as is
the case for the bulk of the waste scheduled for
processing in the SPI Facility. Photofission techniques
are not easily applied to the general fissile assay problem.
If unknown and significant amounts of 2H, ‘Be, 13C, 170,
238U, 23~h, or other isotopes that have low (y,n)
thresholds or significant phototission yields are mixed in
the waste, photofission techniques may indicate false
positive fissile assay amounts, In some very dense
matrices, however, photofission techniques may be
needed. The assay problem is strongly matrix dependent.
Continuing research and development efforts in this field
will likely provide adequate solutions in the next few
years.
B. Active Thermal Neutron Fission Techniques
The nuclear safeguards groups at LASL have de-
veloped18 a z~zcf neu~on source, thermal neutron in-
duced f~sion technique that is currently being specialized
for use with 208-( containers of high density nuclear
waste. The technique, known as the “2S2Cf Shuffler,”
consists of moving a strong (109-1010 n/s) ‘s2Cf neutron
source into an interrogation cavity to induce fission
reactions in any fissile material present and then remov-
ing it to a storage shield during a count cycle. Highly
efticient 3He neutron counters are used to count delayed
neutrons from the induced fissions during this part of the
cycle. In test bed measurements at LASL using a 4.6
x 108 n/s 252Cf source, a 3u-detection limit (for a 700-s
total irradiation/count period) of about 67 mg of 23gPu
was determined. In scaled-up designs of this system
using a 10-25 times stronger 252Cf source and a more
efllcient detection system, plutonium assay detection
limits of about 4 mg have been projected.la A conceptualdesign of the 2s2Cf Shuffler is shown in Fig. 15.
The LASL-EG&G/Santa Barbara waste assay work-
ers are developing techniques has@ in part, on neutron
irradiation using a pulsed photoneutroh source.*9 A pulse
of photoneutrons is thermalized and used to irradiate a
waste container around which is placed a 47t solid angle
neutron detector. Following the irradiation, delayed
neutrons from the induced fission reactions are counted.
In a prototype 208-C barrel-size detection system, a
3cdetection limit of a few milligrams of ‘9Pu or 235U
has been obtained in a typical few 100’s of seconds
irradiation with photoneutrons produced by 10- to
18-MeV bremsstrahlung.
A technique17 originally developed for use in the
Enrichment Plant Safeguards Program at LASL is now
under development for use in the Waste Management
Program. This technique, called the “Differential
Dieaway Technique (DDT)/’ uses pulses from a 14-MeV
neutron generator as a source of ultimately thermalized
neutrons. The generator is placed within a graphite-lined
interrogation cavity (see Fig. 16), which is backed with
polyethylene for additional neutron moderation, re-
flection, and personnel shielding. External to the graphite
layer, but inside the bulk polyethylene layer, are
fast-dieaway moderated 3He neutron detectors. The
neutronic properties of this system are shown in Fig. 17,
which shows the thermalized neutron time history pro-
duced by a pulse from the neutron generator in (a) the
external neutron detector, and (b) the interrogation
cavity with a 208-t’ barrel full of p = 1.0 g/cm3 sand and
vermiculite (SiO~.
The original pulse of neutrons dies out rapidly in the
external detector with no measurable response left after
0.4 to 0.5 ms. On the other hand, thermal neutrons
dieaway in the interrogation cavity with a half-life of
about 0.76 ms. (A lining of cadmium and boron neutron
shielding totally isolates thermal neutrons from the
external detector.) Thus, the interrogation pulse lasts a
long time irt the interrogation cavity and can induce
thermal neutron f~sions in any fissile material present in
the waste barrel. Prompt fission neutrons (hard spec-
trum) from such reactions then propagate to the external
detector where they are recorded.
Figure 18 shows the measured response for a 200-s
interrogation period for samples of varying 233U mass
MU’TROM MKt.DIN~
CC?OR UOOSRATOR
“L!”TnON Oelccronm
\ I-’e “ON’0”’’’0”2&a&a
~ ●OSITION
7UaoSWu.o
M!urnom ● mlez.olnm
NEUTRON •MIRLOIW~ T #- R / wTOP Dc’rccronm
● loe Oc’rccrono
t- ● owou oeTcc?On*
80URCC ORIVC UNIT
.
●
*
●
Fig. 15.
Conceptual design of 208< barrel-size 2$*CfShutller System.
18
U + E TUERMU.nwrRcU .-.
Fig. 16.schematic drawing of the prototype differential dieawaytechnique interrogation cavity and detector.
1 I I I I I 3
I
-T I/2 -laps
●
10’)1
I 1 , , [ II 2 3 4 5 6 7
TIME (ins)
Fig. 17.Pulsed neutron dieaway time data for (a) external detector and(b) interrogation cavity.
* placed within the sand-/vermiculite-filled barrel. A linear
response as a function of 2~sU mass was observed, with
an indicated 3u-detection limit of about 3-5 mg 23SUfor.this prototype system. Additional studies 20have shown a
uniform (to + 100A) interrogation flux profile throughout
the volume of the barrel as well. In other words, this
500 -
g 400-I
SLOPE- 6.2 COUNTS/mg
; 300-
Lz 200 -
100-
‘o= 10 20 30235U4;ASS~mg
60 70 80 90
Fig. 18.Experimental response (100s data collection times) of dif-ferential dieaway technique prototype system as a function of3’$Umass.
technique shows great promise for the assay of fissile
material in low-level, high-density wastes.
Matrices consisting of high-hydrogen-density sludges
or concrete wiU undoubtedly pose technical problems for
some or all of the three techniques discussed above.
Potential corrections to the analysis of this type of data
do exist. If the matrix is known, standards or other
techniques can be used to correct the analysis for the
influence of the matrix. In particular, the DDT method
involves the use of an interrogating flux monitor within
the interrogation cavity. 17 Measurements done to date
indicate that as the interrogating flux is absorbed or
attenuated by waste matrix, the thermal neutron flux
throughout the cavity is proportionately reduced. Thus,
by simply dividing the external detector response by the
internal cavity flux monitor, one effectively obtains an
assay that compensates for matrix effects.
C. Passive Neutron Assay Techniques
The Nuclear Safeguards Program 12s13has long made
extensive use of passive neutron coincidence techniques
to quantify 2413pu content in a great variety of applica-
tions. With a known 240PU isotopic ratio for the material
being assayed, a complete total plutonium mass is simply
calculated. For the proposed use of the SPI Facility,10
the vast majority of the fissile content of the waste to be
prccessed will consist of weapons-grade plutonium
(5-6% 240Pu). In this situation, passive neutron coin-
cidence counting techniques may play a large role.
A well-designed’s 208-( barrel-size passive neutron
coincidence counting system is in the final design stage at
19
LASL. With it (when the system is properly shielded
from the cosmic-ray background), one can obtain a
plutonium assay sensitivity of about 100 mg for 5-6?40
240Pu content material in a 300-s assay.
In addition, work is progressing on a crate-size passive
coincidence detector at LASL.2* When completed and
accompanied with appropriate cosmic-ray shielding, this
detection system will have the capability of assaying
121.92-cm (48-in.) x 121.92-cm (48-in.) x 213.36-cm
(84-in.) crates to an assay sensitivity of about 50-100 mg
plutonium (5-6% 2’OPUcontent) i.n a 1000-s assay period.
Depending on the physical size of the waste containers at
the SPI Facility, one or the other of these passive
systems may be useful.
D. SPI Facility Input screening
Waste screening will prevent many potential criticality
problems. We recommend that it be based on assay
measurements done before the waste enters the final
process stream in the SPI Facility. We also recommend
that such screening use the pulsed neutron generator
DDT system. No matter what size the waste containers
are, an applicable DDT system will be available within a
few years.
This preliminary screening would sort waste into the
broad categories of(1) average TRU content, waste that
has an acceptable fissile content; (2) low TRU content,
waste that is essentially fissile free (to be regarded as an
asset and kept separate for blending purposes); and (3)
high TRU content, waste that has an unacceptably high
fissile content.
At this point the high TRU content waste fraction
should be subjected to a detailed fissile lump analysis,
which may be obtainable with a spatially sensitive
passive neutron measurement proposed as part of the
large crate-size passive neutron detector.z’ An x-ray
analysis combined with one of the passive and/or active
neutron measurements should also reveal the presence of
a fissile lump. If a package containing a high TRU
content contains a fissile lump, special handling and
disposal is required. Under no circumstances should a
lump (2200 g of fissile material in a single piece) be
allowed to enter the SPI Facility. The potential for
criticality, even with the preventative measure of BZ03
poisoning, is too great to think of processing it routinely
through the slagger. The sudden introduction of a large
lump could result in a very rapidly developing criticality
hazard. It is an anomaly and must be processed separate-
ly.
If the high TRU package does not contain lumps of
significant mass, then the package can be blended with
the appropriate volume of the low TRU fraction to
generate an acceptable average TRU mix that can then
be introduced safely into the SPI Facility.
The preliminary screening phase is not a high-quality
assay measurement; t50- 100% quality is adequate.
This can be done very rapidly (<100 s for packages up
to and including crate size) and need not introduce any
significant time delays in the overall processing. A more
accurate (better geometry, better matrix compensation,
etc.) assay measurement for the input/output considera-
tions is best done with the final blended SPI Facility
input material.
IV. ASSAY OF PRODUCT SLAG CASTINGS
If 1.0 x 10-2 wt fraction of BZ03 is added to the SPI
Facility waste input, a fairly accurate assay of the final
91.44-cm-(36 -in.-) diam x 30.48 -cm-( 12-in.-) thick, p =
2.8-3.0 g/cm3, slag casting is obtainable by simple total
neutron counting. (Refer to Table IX.) The known BZ03
content of the slag casting should generate a fairly
constant B(a,n) neutron source on a fixed neutron per
alpha-particle basis. This is based on the fact that thick
target neutron yields are not much different for the
different alpha-particles that exist in the typical TRU
wastes the SPI Facility will be processing (that is, the U,
Pu, and Am isotopes). Thus, the final slag casting
neutroh output will be a quantitative indication of its
total TRU content.
If a total fissile content is desired, we recommend an
active neutron measurement. Considering the expected
uniformity of the final produc~ any of the techniques
discussed earlier could provide this assay. This measure-
ment will be the easiest and by far the most accurate of
all the measurements performed at the SPI Facility. We
do not see a need for slag grab sample assays. The slag
casting can be more than adequately assayed with
presently known techniques.
●
✌
20
V. CONCLUSIONS, RECOMMENDATIONS, AND
COST ESTIMATES
A. Conclusions.
We conclude on the basis of this study that the
proposed SPI Facility will have a minimal criticality
hazard associated with its operation if a three-level
criticality-hazard prevention program is implemented.
The three levels of the program are as follows.
1. A prescreening (50- 100’?loaccuracy, active neutron
assay) of all incoming wastes is performed before they
enter the SPI Facility processing stream. On the basis of
this prescreening into three broad categories (each
category to cover a factor of 3-10 or greater in tissile
material concentration), only an average tissile concen-
tration waste stream is processed by the SPI Facility. No
significant fissile lumps or high-concentration fissile
materials are allowed into the SPI Facility. This prevents
a rapidly developing criticality hazard under most situ-
ations.
2. An approximate 1.0x 10-2 wt fraction of BZ03 (or
equivalent neutron poison) mixture is added uniformly to
all waste. This significantly improves the criticality safety
while not significantly impacting processing volume or
overall SPI Facility economics.
3. Direct criticality-hazard monitors are installed
around the drying, pyrolysis, and slag forming regions
and any other structure of the SPI Facility with a
potential criticality hazard, These monitors should be
based on passive neutron counting using an existing
detector design. An adjunct active neutron system can be
provided for monitoring the build-up of tissile isotopes.
We feel these three independent criticality-hazard
prevention measures will lower the criticality hazard to
an acceptably low level. At the same time, these three
strategies do not add significantly to the cost or etTlcien-
cy of the operation of the SPI Facility.
B. Recommendations
Each of the three measures recommended requires
more detailed study before actual start up of the SPI
Facility. It is our judgment that the facility parameters
(the size and shape of the drying and pyrolysis regions;
composition of waste; detailed incinerator properties
such as temperature and B20J distributions expected;
and many other factors) need to be quantified beyond the
present study. The final system will probably be some-
what different from that used in this study.
C. Cost Estimates
Recommendations of specific instrumentation for the
SPI Facility are based on the current (February 1980)
state-of-the-art technology and on the authors’
projections of what pertinent developments are likely to
occur in the next few years.
1. The direct criticality monitors should be based on
3He/polyethylene open-faced sandwich proportional
counter systems. A separate system will be required for
each of the drying, pyrolysis, and slag regions. This is
based on the MCNP calculational results that showed
neutronic decoupling of these regions. Some develop-
ment of these systems is required, chiefly, experimen-
tal/calculational determination of exact size require
menta, as well as a more detailed study of the false
alarm/optimal sensitivity problem. We estimate that
monitor development costs will be about $200k in 1980
dollars. Equipment costs (these are more uncertain
because exact detector sizes must be determined) for
three separate systems with associated alarming equip
ment will likely be in the $300 -$500k range (1980
dollars). This cost does not include the price of any SPI
Facility system shutdown equipment. The active neutron
addition discussed in the text would add about $50k to
the cost of each system.
2. For the prescreening phase, we recommend an
active neutron system, specifically, the DDT system
because of its more rapid assay capability. We assume
that this system’s development will occur independently
of the SPI Facility needs so that no developmental costs
are included. (If more rapid system development is
required for the SPI Facility, additional funds for this
would be required.) Projected costs for a single large
DDT system [to handle any waste packageof121.92 cm
(48 in.) x 121.92 cm (48 in.) x 213.36 cm (84 in.) or
smaller] will be about $300k, which includes a. spare
neutron generator and pulse transformer as well as all
required personnel shielding. (The basic radiation hazard
of th.s system is not large.) We assume that manual
handling of waste packages with fork lifts or the like can
be used. If fully automated handling is required, addi-
tional costs will be incurred.
21
3. For x-ray and other special measurements required
for the few “problem” waste packages in the prescreen-
ing phase, we recommend a small, low-energy electron
Iinac. Varian Associates of Palo Alto, California, pro-
duce, on a commercial basis, such Iinacs for medical and
quality control x-ray work. A suitable, 8-MeV machine
(with pushbutton 6-,8-, and 10-MeV options) is available
from Varian Associates (linatron 2000) for about $300k
on an operational basis. We recommend installing this in
an outbuilding (a dirt-shielded, large-diameter cement
drain culvert could be used to greatly reduce costs)
rather than incorporating it into the SPI Facility.
4. A large passive neutron coincidence system maybe
desirable as part of the “problem” waste package
measurements. This could be used to independently
determine plutonium mass and possibly to provide lump
detection and location. Because this system would be
intended for use with relatively large plutonium masses
(> 10 g Pu), no elaborate cosmic-ray shielding is re-
quired. We estimate a system cost (no development costs
included) of about $350k.
5. A separate (and smaller) DDT active neutron
system would be required for final product slag casting
assay. This system should cost about $200k.
6. For the highly accurate (5- 10’%o accuracy)
blended waste assay measurement done within the SPI
Facility to provide input/output fissile information @
required), we recommend a separate, specially designed
DDT system. For this purpose, special matrix com-
pensation techniques will probably be required so that a
highly accurate assay can be obtained. We estimate a
$200-$300k development charge and a final design
system cost of about $300k.
ACKNOWLEDGMENT
We thank D. R. Smith and T. McLaughlin of the Los
Alamos Scientific Laboratory for several helpful sugges-
tions and for reviewing the manuscript.
REFERENCES
1. “Use of Borosilicate Glass Raschig Rings as a
2,
22
Neutron Absorber in Solutions of Fissile Material,”
American National Standard, ANSI N16.4 (1971).
LASL Group X-6, W. L. Thompson, Ed.,
‘*MCNP—A General Monte Carlo Code for Neu-
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
tron and Photon Transport,” Los Alamos Scientific
Laboratory report LA-7396-M, Rev. (November
1979). ,
Kaiser Engineers, Inc., “Preliminary Criticality
Evaluation of SPI Gasifier and Secondary Combus-
tion Chamber:’ Idaho National Engineering Labo-
ratory report 79-24-R (April 1979).
Kirk McKinley, EG&G Idaho, Inc., private com-
munication, November 1979.
“Gasifier Design Concept:’ The Ralph M. Parsons
Company project number 961-006-80, drawing
number 61OO6-D- 10-000-01 (September 27, 1979).
“Secondary Combustion Chamber General Ar-
rangement:’ The Ralph M. Parsons Company
project number 961-006-80, drawing number
61006-D-20-000-01 (September 27, 1979).
H. F. Atwater, Los Alamos Scientific Laboratory,
private communication.
F. deHoffman, “Statistical Aspects of Pile Theory,”
The Science and Engineering OJ Nuclear Power,
Vol. 11, C. Goodman, Ed. (Addison-Wesley Publish-
ing Co., Inc., Massachusetts, 1949), pp. 103-119.
J. Orndoff, “Prompt Neutron Periods of Metal
Critical Assemblies,” Nucl. Sci. and Eng. 2, 450
(1957).
S. H. Vegors, Jr., and E. B. Nieschmidt, “Pre-
liminary Investigation of a Criticality Monitoring
Technique for a Transuranic Waste Incinerator;’
Idaho National Engineering Laboratory report
TREE- 1285 (October 1978).
R. H. Augustson and T. D. Reilly, “Fundamentals
of Passive Nondestructive Assay of Fissionable
Material,” Los Akimos Scientific Laboratory report
LA-565 1-M (September 1974).
N. Ensslin, M. L. Evans, H. O. Menlove, and J. E.
Swanson, “Neutron Coincidence Counters for
Plutonium Measurements,” J. Inst. Nucl. Mater.
Manage, INMM VII, 43 (Summer 1978).
.
.
13.
14.
15.
16.
K. P, Lambert and J. W. Leake, “A Comparison of
the V.D.C. and Shift Register Neutron Coincidence
Systems for 24’WUAssay,” J. Inst. Nucl. Mater.
Manage, INMM VII, 87 (Winter 1978-1979).
E. J. Dowdy, C. N. Henry, A, A. Robba, and J. C.
Pratt, “New Neutron Correlation Measurement
Technique for Special Nuclear Material Assay and
Accountability,” Proceedings of the International
Symposium on Nuclear Material Safeguards, Vien-
na, Austria, October 2-6, 1978 (IAEA, Vienna, May
1979), Vo]. II, pp. 125-136.
C. D, Ethndge, E. J. Dowdy, C, N, Henry, and D.
R. Millegan, “A Microprocessor Based Neutron
Count Moment Logic Module for Special Nuclear
Material Assay by the Neutron Fluctuation Meth-
od:’ Proceedings ESARDA First Annual Sym-
posium on Safeguards and Nuclear Material Man-
agement, Brussels, Belgium, April 25-26, 1979
(Brussels, 1979), pp. 320-324.
E. J. Dowdy, G. E. Hansen, A, A. Robba, and J. C.
Pratt, “Effect of (a,n) Contaminants and Sample
Multiplication on Statistical Neutron Correlation
Measurements,” Proceedings ESARDA Second An-
nual Symposium on Safeguards and Nuclear Materi-
al Management, L. Stanchi, Ed,, Edinburgh,
Scotland, March 26-28, 1980 (Edinburgh, 1980), pp.
359-363.
17.
18.
19.
20.
21.
W. E. Kunz, J. D. Atencio, and J. T. Caldwell, “A
l-nCi/g Sensitivity Transuranic Waste Assay Sys-
tem Using Pulsed Neutron Interrogation,” submitted
to the Twenty-fwst Annual Meeting of the INMM,
Palm Beach, Florida, June 30, July 1-2, 1980.
T. W. Crane, “Measurement of Pu Contamination
at the 10-nCi/g Level in 55-Gallon Barrels of SolidWxte with a z~2Cf Assay System,” Proceedings of
the International Meeting on Monitoring of
Pu-Contaminated Waste, J. Ley, Ed., Ispr& Italy,
September 25-28, 1979 (Ispra, 1980), pp. 217-226.
M. R. Cates, “Application of Linear Accelerator
Technology to the Detection of Trace Amounts of
Transuranics in Waste Barrels,” submitted to the
Symposium on the Management of Alpha-
Contaminated Wastes, Vienna, Austria, June 2-6,
1980.
G. Robert Keepin, Ed., “Nuclear Safeguards Re-
search and Development Program Status Report,
Octolxr-December 1979~’ Los Alamos Scientific
Laboratory report LA-824 1-PR (May 1980).
J. T. Caldwell, “High Density Waste Assay: Large
Crate Counting via Passive Neutron Counting,”
submitted to the Symposium on the Management of
Alpha-Contaminated Wastes, Vienna, Austria, June
2-6, 1980.
APPENDIX
SPI FACILITY GEOMETRY FOR MCNP
The MCNP code requires an explicit description of all Figs. A-20—A-3 1. The small numbers appearing on the
surfaces in the geometrical model. This appendix sup- tigures are surface numbers as required by MCNP to
plies this information for the SPI Facility, excluding all describe the geometry. Table A- I gives a complete listing
secondary structures. Figures A- l—A-31 are a com- of all the surfaces needed to describe the geometry.
prehensive set showing the incinerator in the yz, XZ, and These are included for completeness and for the reader
xy planes. Cuts in the yz plane at important x positions who wants to know the details of the geometry. These
are shown in Figs. A-1 —A-6; cuts in the xz plane at surfaces can be interpreted by referring to Ref. 2.
important y positions are shown in Figs. A-7—A- 19;
cuts in the xy plane at important z positions are shown in
23
z1-Y
X*O
I
8
L
717
,23 23 16
25.
9
1
14’ 3+-4 4+-3
Fu. A-L
View of the geometrkal model cf the SPI for the MCNPcalculations.This is a cut in the yz plane at the point x = Ocm.
.2 .1
1- 2“
18
l-f
8
7z )7
Y23 23 ,16
x=55
H /9 w “- \ IIII v \ II
14’ 3+-4 4-+-3
Fig. A-3.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the yz plane at the point x = 55cm.
z1-Y
X=loo
view of thecalculations.
zLY
X=30
1- .2 2, 4
18
8
7
,17
k 23 23+1 I /16
,9
14’ 3Q-4 4+3
Fig. A-Z
View of the geometrical model of the SP1 for the MCNPcalculations. This is a cut in the yz plane at the point x = 30cm.
z1-Y
.X=70
1- 2\
,2 -1
18
8
717
23 23
?
Ih ,10 Is-i ,11 24 6
=30
Fig. AA.View of the geometrical mo&l of the SPI for the MCNPcsstculations. This is a cut in the yz plane at the point x = 70cm.
la
25it’wkr
Fig. A-5.geometrical model d the SPI for the MCNPThis is a cut in the yz plane at the point x = 100
cm.
24
18
.
1,
z
18
z
+ 17
x =200 ,16
t
29
14\
Fig. A-6.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the yz plane at the point x = 200cm.
18/
z1-x
Y=30
T’-rI 9
I .14I 1 /
Fig. A-8.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xz plane at the point y = 30
z1-x
Y=121.92
.
+ 17,
Y=o /23 16
14
Fig. A-7.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xz plane at the pointy = Ocm.
z
--x--Y=48
1- Oi
,18
8
7,17
23 2316
I I/
106’
,6
,9
’14
Fig. A-9.
View of the geometrical mo&l of the SPI for the MCNPcalculations. This is a cut in the xz plane at the point y = 48cm.
1, /2
8
7 ,17
L,23 23’
16
10
6’/6
9
’14
Fig. A- 10.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xz plane at the point y =121.92 cm.
25
I
z1-x
Y=168,91
1. I2“
-2
18
8
7
23---1 1’23
,16
I I
’14
Fv. A- 11.
View of the geometrical model of the SPI for the MCNPcalculations. Tlris is a cut in the xz plane sst the point y =168.91 cm.
—x
Y=21590
View of thecalculations.215.90 cm.
18
/}1
6, 21 22 10 -6
Fig. A- 13.geometrical mcdel of the SPI for the MCNPThis is a cut in the xz plane at the point y =
*
Y= 235.52
z1-x
Y = 194
2 2
N /–1
1-
,18
8,17
I
1s .1
t-3 ‘ 14
Fig.A-12.
View of the geometrical model of the SP1 for the MCNPcalculations. This is a cut in the xz plane at the pointy = 194cm.
18
tzL- 8
!7x
Yx230.6216
2323
Fig. A. 14.
View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xz plane at the point y =
23i).62 cm.
,18
II
,66- 10
22
1421’
3 II-4 4+3
Fii. A-15.
View of the geometrical model of the SPI for the MCNP
.
.
calculations. This is a cut in the xz plane at the point y =235.52cm.
●
✎
1-18
z
x17
Y=284.37 .16
6- 11 /6
10 10
21. ’22
3-+4 4. 3\
14
Fig. A- 16.
zLx
Y = 300
,17
16
6\ ..6
14
Fig. A- 17.View of the geomctricsd model of the SP1 for the MCNP View of the geometrical m-odel of the SP1 for the MCNPcnkuiadons. This is a cut in the xz planeat the point y = calculations. This is a cut in the xz plane at the pointy = 300284.37 cm. cm.
L la
z
x
Y=33017
16
6, 6
14
34 I-3
Fig. A-18.
View of the geometrical model of the SP1 for the MCNPcalculations. This is a cut in the xz plane at the pointy = 330cm.
‘L- 0x
Z.-3OO
Fig. A-20.View of the geometrical model of the SP1 for the MCNPcalculations. This is a cut in the xy plane at the point z = –300cm.
,18
zL 17
x ,16
Y =337 I I
--!-t_JFig. A-19.
View of the geometrical model of the SPI for the. MCNPcakulations. This is a cut in the xz plane at the pointy = 337cm.
Y
Lx
Z.-16
Fig. A-21.View of the geometrical model of the SP1 for the MCNPcalculations. This is a cut in the xy plane at the point z = – 16cm.
27
29 29/
Y
Lx
Z=-1
/
Fig. A-22.View of the geometrical model of the SPI for the MCNPcalculations. Thii is a cut in the xy plane at the point z = –1cm.
Y
Lx
ZX52
29
/6
25
Fig. A-24.View of the geometrkal model of the SPI for the MCNPcrdculations. This is a cut in the xy plane at the point z = 52cm.
Y
Lx
2=0
L-LFig. A-23.
Vkw of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z = Ocm.
Y
Lx
z~53.3
29
6
25
Fig. A-25.
view of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z = 53.3cm.
*
.
.
28
w
.
Y
Lx
Z=71.806
’25
Fig. A-26.View of the geometrical model of the SPI for the MCNPcalculations.TIds is a eut in the xy plane at the point z = 71.80cm.
Y
Lx
Z =116.45
29
Fig. A-28.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z =
116.45 cm.
Y 1-
07
x 8
Z .243
Fig. A-30.View of the geometrical mcdel of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z = 243cm.
u 15
Y 6
x
Z=I16
\ 25
Fig. A-27.
View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z = 116cm.
Y
L
23
x
Z =160
Fig. A-29.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z = 160
cm.
Y
~
a’2
x
2.300
Fig. A-31.View of the geometrical model of the SPI for the MCNPcalculations. This is a cut in the xy plane at the point z = 300cm.
29
TABLE A-I
GEOMETRY OF INCINERATOR
1 C/Z O 121.9281.28
2 C/Z O 121.9273.66
3 C/Z O 261.6268.58
4 C/Z O 261.6266.04
5 C/Z O261.6260.966 C/Z O 168.91168.917 K/Z O 121.92365.9605263 0.3638196019E- 18 K/Z O 121.92378.5936842 0.3638196019E-1
9PZ010 P O 2.300980615E-3 8.587376556E-3 1
11 P O2300980615E-3 8.587376556E-31.2709765391
12P O 1.732050806E-1354.6052253
13 P O 3.366631827E-35.936280184E-4 1
14 PZ –15.24
15 PY 215.90 ,16 PZ 116.45
17 PZ 163.83
18 PZ 243.84
19PZ 612.14
20 Pz 868.6821 PX –15.24
22 Px 15.2423 C/Z O 121.92121.92
24 PY 284.3710495
25 PY O
26 PZ –106.68
27 PZ –191.44
28 PZ –28288
29 PY 337.82
30 PY 193.04
31 C/Z O 261.623000
32 PZ –100.00
33 so 130000m
34 PZ 335.27
●
&
30
Domestic NTlsPa.seRange Price Price Code
001422s s S.oo A02026+150 6.00 A0305147s 7.00 A04
076-100 8.00 AO$101-125 9.00 A06116-1 SO 10.00 A07
printed in the United Slates of America
Available fromNational Technical Information service
US Department of CommerceS28S PorI Royal RoadSpringfield, VA 22161
Microfiche $3.S0 (AO1)
Domestic NTIS Domestic NTIS
Page Range Rice Rice Code Page Range Rice pfiCC Code
151-17s S1l.00 A08 301-32s S17.00 A14
176-200 12.00 A09 326-350 18.04 A15
201-22s 13.00 AIO 351 -37s 19.00 A16
226-250 14.00 All 376400 20.00 A172S1-17S 1s.00 A12 401425 21.00 A18
276-300 16.00 A13 4264s0 22.00 A19
Domestic N-I-ISPage Range Rice Price Code
4s147s S23.00 A20476-500 24.00 A21
SO1.S25 2s.00 A22S26-S50 26.oo A23551-575 27.00 A24576400 28.00 A2560! -Up t A99
tAdd S1.00 for each addi@ul 2S-page increment or portion thereof from 601 pages up.
●
e>“
Top Related